I think the author was too pessimistic. He draws a direct inverse relationship between AI-use and IQ, but in my experience it's kind of U-shaped. The incompetent people use it as a crutch (mostly failing students in my experience), but competent people use it well to accelerate their productivity.

There's definitely an upstream effect on student learning, and it may be producing poor graduates/juniors. However, I reckon that's growing pains as teachers learn to adopt their curriculums to better account for AI.

I think about it like calculators. Calculators probably reduced our ability to do mental math, and it made purely arithmetic take-home assignments pointless. But it also opened students up to learn much more advanced mathematics quickly, allowing people to attain a high degree of mathematical productivity much earlier than they otherwise would have.